Political Psychology, Vol. 21, No.

1, 2000

On the Relationship Between Attitude Involvement

and Attitude Accessibility: Toward a
Cognitive-Motivational Model of Political
Information Processing
Howard Lavine
State University of New York at Stony Brook

Eugene Borgida and John L. Sullivan

University of Minnesota

A model of the relationship between attitude involvement and attitude accessibility was
developed and tested. The model specifies that attitude involvement leads to selective
(biased) issue-related information-gathering strategies, which in turn produce extreme and
univalent (unambivalent) attitudes. Finally, attitudes associated with univalent and extreme
underlying structures should occasion relatively little decision conflict and thus should be
highly accessible. Questionnaire response data gathered in a national telephone survey and
from two samples of undergraduates revealed that both attitude extremity and attitude
ambivalence on selected political issues mediated the relationship between attitude
involvement and attitude accessibility. Some findings indicated that selective processing
mediated the relationship between attitude involvement and attitude extremity and
ambivalence. Discussion focuses on the processes linking involvement to accessibility, the
factors that moderate the ambivalence-accessibility relationship, and the relevance of the
model to media-based priming effects and to the nature of public opinion and the survey
response.
KEY WORDS: attitude accessibility, attitude involvement, political information processing.

Increasingly, political psychologists are devoting attention to the cognitive

mechanisms that underlie political behavior. As a result, new theoretical perspec-
tives and methodological techniques that focus on cognitive structure and process
have been developed (or appropriated from allied disciplines, primarily cognitive
psychology; see, e.g., Meyer & Schvaneveldt, 1971; Neisser, 1976). Investigators

81
0162-895X © 2000 International Society of Political Psychology
Published by Blackwell Publishers, 350 Main Street, Malden, MA 02148, USA, and 108 Cowley Road, Oxford, OX4 1JF, UK.
82 Lavine et al.

using cognitive perspectives have already produced valuable insights into several
key topics of interest to political psychologists, including opinion measurement
(e.g., Zaller & Feldman, 1992), candidate appraisal (e.g., Fazio & Williams, 1986;
Lau, 1989; Lodge, McGraw, & Stroh, 1989; Rahn, Aldrich, Borgida, & Sullivan,
1990), and the organization of political belief systems (Judd, Drake, Downing, &
Krosnick, 1991; Lavine, Thomsen, & Gonzales, 1997).
Response latency, or reaction time, is among the most promising and widely
used methodological innovations for exploring the cognitive underpinnings of
political behavior and judgment (e.g., Bassili, 1996; Fazio & Williams, 1986; for
a review, see Lavine, 1997). Reaction time has been used to measure various
aspects of cognitive functioning, such as construct accessibility and spreading
activation, and it has contributed importantly to our understanding of politically
relevant social phenomena such as impression formation (Smith & Miller, 1983),
stereotyping and prejudice (Dovidio, Evans, & Tyler, 1986), and the relationship
between attitudes and behavior (Fazio, Chen, McDonel, & Sherman, 1982).
Within political psychology, response latency has been most often used to
assess the construct of attitude accessibility (e.g., Bassili, 1995, 1996, 1998; Fazio
& Williams, 1986; Krosnick, 1989; Lavine, Sullivan, Borgida, & Thomsen, 1996).
Attitude accessibility refers to the ease and quickness with which a person can
retrieve an attitude from memory and use that attitude in making a judgment or
decision (for a theoretical review, see Fazio, 1986). Research in social and political
cognition indicates that a given construct (e.g., a personality trait, a government
policy attitude) must be both available in memory and cognitively accessible to
guide information processing and influence judgment and decision-making
(Aldrich, Sullivan, & Borgida, 1989; Higgins & King, 1981). A number of studies
in the political realm have provided strong evidence that accessible attitudes exert
more powerful effects on judgment and decision-making than do relatively inac-
cessible attitudes. For example, Fazio and Williams (1986) found that voters with
highly accessible attitudes toward Reagan and Mondale (i.e., voters who could
respond quickly to inquiries about their attitudes toward the candidates) were more
likely to vote in the 1984 presidential election in a manner consistent with those
attitudes than were respondents with less accessible attitudes toward the candidates
(see also Bassili, 1995). In related research, Lau (1989) assessed the accessibility
of voters’ candidate, issue, group, and party schemas by counting the frequency
with which voters offered information from each of these categories (in an
open-ended interview) for liking and disliking presidential candidates and political
parties. Lau showed that the accessibility of these schemas was stable over 4-year
periods and that such schemas exerted a greater impact on candidate evaluation
and voting behavior when they were highly accessible.
An important conclusion of these studies is that political attitudes and other
relevant constructs (e.g., personality traits, party identifications) are likely to guide
political judgments such as candidate evaluation (e.g., Bassili, 1998; Fazio &
Williams, 1986; Lavine et al., 1996) and presidential performance (e.g., Iyengar &
Involvement and Accessibility 83

Kinder, 1987; Iyengar, Kinder, Peters, & Krosnick, 1984) to a greater extent when
they are easily retrievable from memory—that is, when they are highly accessible.
What is less well understood is how attitudes become cognitively accessible. That
is, relatively little theory and research in political psychology has been devoted to
the cognitive and motivational origins of attitude accessibility. This is an important
theoretical question for political psychologists, for if we can better understand how
and why certain political ideas come to dominate the belief systems of individu-
als—that is, if we can gain some insight into the origins of attitude accessibil-
ity—we stand to deepen our understanding of at least some of the psychological
roots of mass political behavior. What factors lead people to become involved in
political issues, and how does involvement lead to attitude accessibility?

From Attitude Involvement to Attitude Accessibility: A

Cognitive-Motivational Model

According to construct accessibility theory (Higgins & King, 1981), an atti-

tude’s accessibility is determined by the frequency and recency with which it is
thought about or expressed, its relation to people’s ongoing needs and goals, and
the extent to which the attitude is linked in memory to other constructs (e.g., beliefs,
attitudes, values). As Krosnick (1989) and Lavine et al. (1996) have noted, each of
these attributes is associated with attitudes marked by high levels of involvement.
Highly involving attitudes are closely linked with people’s tangible goals, core
values, and socially important individuals and groups (Boninger, Krosnick, &
Berent, 1995), and thus are the targets of frequent conscious thought. In turn,
attitude-relevant thought strengthens the association in memory between an atti-
tude object and its evaluation, thereby heightening attitude accessibility (Fazio et
al., 1982). Highly involving attitudes are also more likely to be embedded in a larger
structure of other attitudes, beliefs, and values (Judd & Krosnick, 1989; Lavine,
1994; Lavine et al., 1997; Thomsen, Lavine, & Kounios, 1996), thus promoting
accessibility indirectly through the direct activation of linked constructs.
In our view (Thomsen, Borgida, & Lavine, 1995; see also Boninger et al.,
1995), a political issue should be important or involving to a person to the extent
that it impinges on, reflects, or is otherwise associated with the self. When an issue
or event implicates the self, involvement in that issue (i.e., the subjective perception
that the issue is personally important) should be aroused. A political issue or event
can implicate a variety of aspects of the self-concept. Perhaps the earliest discussion
of the aspects of the self-concept is William James’ (1890) tripartite theory. James
argued that the self includes material aspects (e.g., one’s body, material posses-
sions), social aspects (e.g., one’s spouse, children), and spiritual aspects (e.g., one’s
values, religious and political beliefs). Political issues or events that impinge on
any of these dimensions of the self may occasion involvement. Indeed, recent
attempts to explicate the origins of attitude involvement have resulted in taxono-
mies that can be mapped onto James’ tripartite model (Boninger et al., 1995; Fiske
84 Lavine et al.

& Neuberg, 1990; Greenwald & Breckler, 1985; Johnson & Eagly, 1989; Katz,
1960; Smith, Bruner, & White, 1956; for a review, see Thomsen et al., 1995).
Accordingly, a person should deeply care about a political issue to the extent that
it is seen as relevant to his or her tangible or material goals (self-interest), to the
extent that people and groups that are important to the person are seen as caring
about or affected by the issue (social identification), and to the extent that the issue
is viewed as being linked to the attainment of cherished values (value relevance).
Recent studies have provided strong evidence that self-interest, social identifica-
tion, and value relevance independently predict attitude involvement (Boninger et
al., 1995; Thomsen et al., 1995).
We argue that the degree of involvement aroused in an issue is an important
determinant of the type of motivation that characterizes the individual’s subsequent
information processing related to that issue (Figure 1). When message recipients
first encounter an issue that they perceive to be relevant to their self-concepts,
motivation to process information about that issue should increase. Moreover,
because recipients know little about the issue and its potential consequences, they
are likely to adopt an objective or “validity-seeking” approach to gaining informa-
tion about the issue (e.g., Chaiken, Liberman, & Eagly, 1989; Petty & Cacioppo,
1986). That is, people should be initially equally receptive to positive and negative
information about the issue in an effort to develop an attitude that “squares with
the relevant facts” (Petty & Cacioppo, 1986). This reasoning is consistent with the

Figure 1. A model of the involvement-accessibility relationship.

Involvement and Accessibility 85

findings of persuasion research conducted within the dual-process frameworks

(Chaiken et al., 1989; Petty & Cacioppo, 1986).1
However, as involving attitudes solidify and crystallize over time, they should
become positively associated with issue-relevant knowledge and with the motiva-
tion for attitudinal defense (e.g., McGraw, Fischle, Stenner, & Lodge, 1996).
Issue-related information-seeking strategies are thus likely to shift from being
relatively objective at the outset to being biased in favor of congenial information
(i.e., information that supports the person’s attitude) once the attitude is crystal-
lized. That is, as attitudes become increasingly connected with the self—as they
become more closely linked with our plans, goals, and cherished values—we
should be more likely to experience their negative implications as highly threaten-
ing, and thus should be more motivated to refute or counterargue (or go out of our
way to avoid) such incongruent information (for reviews, see Frey, 1986). For
example, a woman who cares a great deal about the issue of legalized abortion and
has a positive attitude toward the issue is likely to be motivated to “defend” her
pro-legalization attitude rather than objectively integrate the positive and negative
implications of her position in favor of abortion rights (for a discussion of validity-
seeking and defense motivation within and beyond the persuasion context, see
Chaiken et al., 1989; see also Eagly & Chaiken, 1998). This reasoning is consistent
with persuasion theory and research showing that with respect to prior attitudes
(i.e., those existing before the experiment), involvement produces both resistance
to change (e.g., Gorn, 1975; Rhine & Severance, 1970; Zuwerink & Devine, 1996;
for a review, see Johnson & Eagly, 1989) and biased message processing (e.g.,
Borgida & Howard-Pitney, 1983; Howard-Pitney, Borgida, & Omoto, 1986; Sherif
& Hovland, 1961).
To the extent that individuals are motivated to defend their attitudes rather than
objectively evaluate them, how might they accomplish this goal? At least two
selective information-seeking strategies are available (for a review, see Eagly &
Chaiken, 1993). First, when highly involving attitudes are at stake, attitudinal
defense can be achieved by selectively attending to and processing information that
confirms rather than challenges the validity of one’s opinions (e.g., Festinger, 1964;
for a review, see Frey, 1986). For example, a person with a committed position on
the death penalty should—all else equal (e.g., Freedman & Sears, 1965)—prefer
to read about a study that supports rather than refutes his or her beliefs about the
deterrence effect of the policy. Theoretically, selective exposure occurs as people
attempt to reduce cognitive dissonance associated with the acceptance of incon-
gruent information (Festinger, 1957, 1964). Although early research (summarized

1 This “validity-seeking” orientation is conceptually similar to the predecisional openmindedness of

information seeking noted by cognitive dissonance theorists (Brehm & Cohen, 1962; Festinger, 1957,
1964). Selective (dissonance-reducing) information seeking is hypothesized to occur only after
decisions are enacted (see Frey, 1986).
86 Lavine et al.

by Freedman & Sears, 1965) was largely unfavorable to the selective exposure
hypothesis, more recent studies (summarized by Frey, 1986) suggest that a prefer-
ence for exposure to congenial information reliably occurs when people are highly
committed to their attitudes and decisions (i.e., when involvement is high; e.g.,
Brock & Balloun, 1967; Frey & Stahlberg, 1986; Schwarz, Frey, & Kumpf, 1980).
For example, Pomerantz, Chaiken, and Tordesillas (1995) found that a measure of
attitude “embeddedness” (a factor representing an issue’s centrality to the self,
personal importance, value relevance, and the respondent’s issue-relevant knowl-
edge) on the issue of capital punishment predicted greater levels of selective
exposure to attitude-congruent versus incongruent information (measured by sur-
reptitiously timing respondents’ attention to congruent vs. incongruent information).
Attitudes can also be defended by cognitively responding to issue-relevant
information in a selectively critical manner. Specifically, information that refutes
preferred positions can be subject to greater scrutiny and counterargumentation
than information that supports preferred positions. In a classic study of selective
judgment, Lord, Ross, and Lepper (1979) presented participants who held extreme
attitudes for or against capital punishment with two ostensible empirical reports on
the efficacy of the policy as a deterrent to murder. One report provided evidence
in support of capital punishment; the other suggested that capital punishment leads
to more murders. Lord et al. found a “biased assimilation” effect on participants’
judgments of the perceived quality of the empirical studies: Respondents with
positive attitudes toward capital punishment found the study supporting the effi-
cacy of the policy to be more convincing and better conducted than the study
opposing the efficacy of capital punishment, and the reverse was true for respon-
dents with negative attitudes toward capital punishment (see also Houston & Fazio,
1989; Sherif & Hovland, 1961).2 Moreover, Lord et al.’s (1979) participants
reported that their attitudes polarized (i.e., became more extreme) after reading the
two empirical studies. Research supports the notion that selective processing (e.g.,
of exposure or judgment) in the defense of prior attitudes is associated with attitudes
that have “strength-related” properties such as involvement (e.g., Houston & Fazio,
1989; Lord et al., 1979; Pomerantz et al., 1995; Sherif & Hovland, 1961).
The nature of information processing—whether it is validity-seeking or defen-
sive—should in turn determine the mix of considerations (i.e., reasons for support-
ing or opposing a policy; Zaller & Feldman, 1992) underlying the attitude. When
the processing motivation is objective, people are likely to be exposed to and to
accept a mix of issue-related considerations that support opposing sides of a policy
debate. For example, a person might simultaneously believe that capital punish-
ment is morally wrong but that it is effective in deterring violent crime. Under these

2 Attitudinal defense may also be facilitated by selective recall of information that comports rather than
conflicts with one’s opinions. However, the selectivity effects of recall appear less consistent than
those associated with exposure or judgment (see Eagly, Chen, Chaiken, & Shaw-Barnes, 1999).
Involvement and Accessibility 87

processing conditions, the attitude’s supporting belief structure should exhibit (at
least some) evaluative inconsistency. In contrast, when people have directional
goals—that is, when they are motivated to defend the validity of preexisting
attitudes—they are likely to be predominantly exposed to and to accept congenial
considerations. Thus, under biased processing conditions, the attitude’s underlying
structure of feelings and beliefs should exhibit relatively high levels of evaluative
consistency.
The recognition that people often simultaneously hold positive and negative
beliefs about a political issue or candidate suggests a different way of understanding
how such attitudes are represented in the cognitive system. Until recently, attitude
theorists have almost universally made the implicit assumption that opinions are
represented in terms of a single, bipolar (positive-negative) dimension (for a
review, see Eagly & Chaiken, 1993). Increasingly, however, investigators have
acknowledged that many attitudes are characterized by the coexistence of both
positive and negative evaluations (e.g., Alvarez & Brehm, 1995; Cacioppo &
Berntson, 1994; Cacioppo, Gardner, & Berntson, 1997; Feldman & Zaller, 1992;
Hochschild, 1981; Lavine, Huff, Wagner, & Sweeney, 1998; Lavine, Thomsen,
Zanna, & Borgida, 1998; Thompson, Zanna, & Griffin, 1995; Zaller & Feldman,
1992). This suggests that attitudes are instead represented in terms of two separate
unipolar dimensions (i.e., one negative and one positive).
Variation in information-processing goals should thus have two key structural
consequences relevant to the unipolar, bidimensional view of attitudes. First, the
use of selective processing should result in attitudes with univalent structures, or
structures in which positive or negative evaluations of the issue—but not both—are
present (or strong). By contrast, unselective processing should produce attitudes
with relatively ambivalent structures, or those in which both positive and negative
evaluations are present. Second, relative to objective processing, selective process-
ing should produce attitudes that are evaluatively extreme. Recent research sug-
gests that attitudes are based on the canvassing and integration of accessible
considerations or beliefs about the issue in long-term memory (Lavine, Huff, et al.,
1998; Tourangeau & Rasinski, 1988; Zaller & Feldman, 1992). To the extent that
the considerations are evaluatively similar, they should integrate to produce ex-
treme overall attitudes. However, when the underlying base of considerations is
evaluatively mixed, integration of the opposing implications should produce rela-
tively moderate attitudes (see Anderson, 1971; Fishbein & Ajzen, 1975; Tetlock,
1986).
To simplify: Attitudes that receive selective processing should result in struc-
tures in which the attitude object (e.g., a political issue) is strongly linked with one
type of evaluation (positive or negative), but not both. We refer to such univalent
attitudes as single-evaluation attitude structures. In contrast, attitudes that are not
the targets of selective processing are likely to develop links to both types of
evaluation, possessing what we refer to as dual-evaluation attitude structures.
88 Lavine et al.

Whether an attitude can be characterized by a single- or a dual-evaluation

structure has direct implications for that attitude’s accessibility in memory.
Specifically, with respect to attitudes that have dual-evaluation structures,
responding to a speeded attitudinal inquiry (i.e., a response latency task) should
require the integration of positive and negative evaluative components. In other
words, the retrieval of ambivalent or dual-evaluation attitudes is likely to be
met with at least a modicum of decision conflict based on the underlying
evaluative inconsistency. The extra time required to carry out this information
integration or inconsistency resolution task should cause these attitudes to be
less cognitively accessible than their univalent or single-evaluation counter-
parts. In contrast, attitudes characterized by single-evaluation structures should
occasion relatively little decision conflict, and are thus more likely to be directly
retrieved from memory.
In sum, we propose that involvement in a political issue is based on the degree
to which the issue has implications for one’s goals and plans or the facilitation of
one’s core values, or is of concern to individuals and social groups with which one
strongly identifies (see also Boninger et al., 1995). If involvement is sufficiently
aroused, one should be motivated to defend one’s attitude (rather than to objectively
seek the most “valid” attitude) and to selectively seek (or judge) information about
the issue that affirms the correctness or efficacy of one’s position (and to avoid or
counterargue information that disaffirms one’s position). These selective informa-
tion-gathering strategies should produce attitudes with univalent (e.g., single-
evaluation) structures. Such univalent attitudes are characterized by a relative
absence of decision conflict (and clear behavioral cues), and should thus be highly
accessible in memory.

The Present Studies

To provide a preliminary test of the involvement-accessibility model, we

conducted a series of surveys. In study 1, we used national survey data to
examine the origins of attitude accessibility toward the issue of women’s rights.
In study 2, we used a college student sample to assess attitudes toward 14 po-
litical issues, and examined whether respondents held more extreme, less
ambivalent, and more accessible attitudes toward issues that they rated as most
important than toward issues that they rated as less important. In study 3, we
used a college student sample to examine whether the relationship between
involvement (in the issue of affirmative action) on the one hand, and attitude
extremity and ambivalence on the other, is mediated by selective exposure to
attitude-consistent information.
Involvement and Accessibility 89

Study 1

The survey was a national random-digit computer-assisted telephone inter-

view.3 The survey population was defined as all English-speaking adults (18 years
of age or older) residing in households with telephones within the 48 contiguous
states. The number of completed surveys was 1,464; the response rate was 65.5%.4
The study involved investigators from several different institutions and was fo-
cused on attitudes and beliefs about a wide variety of social and political issues.
Each team contributed its own context (e.g., question order) experiment within the
survey. Our portion of the survey consisted of items pertaining to the issue of
women’s rights and assessed the constructs of self-interest, value relevance, social
identification, attitude involvement, attitude extremity, attitude ambivalence, and
attitude accessibility.

Measurement of Constructs

Self-interest. Consistent with previous work (e.g., Crano, 1995), we defined

self-interest as an individual’s perception of the extent to which a political policy
has direct implications for his or her goals, plans, or tangible outcomes. Specifi-
cally, using a 4-point scale (1 = a great deal, 3 = some, 5 = a little, 7 = not at all),
we asked, “How much have policies related to women’s rights directly affected
your life?” Responses were reverse-coded so that higher self-interest scores re-
flected stronger perceptions of self-interest in the women’s rights issue.
Value relevance. Value relevance is defined as the extent to which core values
are viewed as being relevant to the policy issue. Specifically, we used 4-point scales
(1 = very related, 3 = somewhat related, 5 = slightly related, 7 = not at all related)
to assess the extent to which the values of equality and self-respect (both pretested
to be at least somewhat relevant to the women’s rights issue) were related to
respondents’ attitudes toward women’s rights. Responses were reverse-coded and
averaged so that higher value-relevance scores reflected greater levels of value
relevance.
Social identification. Social identification is defined as the extent to which a
person perceives that groups with which he or she identifies regard the issue to be

3 The survey was funded by a National Science Foundation grant to Paul Sniderman, Henry Brady, and
Phil Tetlock, and administered by the Survey Research Center at the University of California at
Berkeley.
4 Consistent with survey research practice, weights were created for each respondent to compensate for
differences in probabilities of selection and to adjust the sample to match certain demographic
distributions. The weighting procedure adjusted for seven variables: gender, race, age, education,
number of eligible adults in the home, number of phone lines in the home, and whether the selected
household had a listed or an unlisted telephone number (the survey greatly oversampled homes with
listed numbers—by a ratio greater than 9). The weighting procedures did not alter the pattern of means
or alter the significance of any of our findings.
90 Lavine et al.

important. We assessed group identification by asking respondents to rate the

extent to which Catholics, Democrats, gays and lesbians, and women in general
(1 = very important, 3 = somewhat important, 5 = not very important, 7 = not at all
important) perceive the issue of women’s rights to be important. All groups were
pretested as being viewed as caring at least slightly about the women’s rights issue.
Respondents also completed 101-point feeling thermometer ratings of each group.
Social identification scores were computed by multiplying the group importance
rating by the thermometer rating for each group, and averaging the four products.5
Attitude involvement. Attitude involvement is defined as the extent to which
a person attaches subjective importance to a given issue (see Krosnick, Boninger,
Chuang, Berent, & Carnot, 1993; Lavine, Huff, et al., 1998). Consistent with
previous work (e.g., Krosnick et al., 1993; Lavine, Huff, et al., 1998), we asked
respondents to rate the extent to which the issue of women’s rights is personally
important and the extent to which it is important that the government does what
the respondent thinks is best on issues related to women’s rights (1 = very impor-
tant, 3 = somewhat important, 5 = not too important, 7 = not at all important).
Responses were reverse-coded and averaged so that higher scores reflected higher
levels of attitude involvement in the women’s rights issue.
Attitude extremity. Attitude extremity refers to the extent to which a person’s
attitude toward a given issue deviates from the midpoint of the scale in either the
negative or positive direction. To assess extremity, we computed the absolute
deviation from a 5-point semantic differential attitude scale toward women’s rights,
where 1 = bad and 5 = good. Scores could thus range from 0 to 2, with higher scores
reflecting more extreme attitudes.
Attitude ambivalence. Attitude ambivalence refers to the extent to which a
person’s attitude toward a given issue contains both positive and negative thoughts
or feelings (Lavine, Thomsen, et al., 1998; Priester & Petty, 1996; Thompson et
al., 1995). Ambivalence can be assessed either “objectively” by integrating sepa-
rate unipolar assessments of positive and negative components of the attitude, or
subjectively by asking the respondent to report the subjective perception of having
“mixed” feelings or beliefs about the issue (see Priester & Petty, 1996; Thompson
et al., 1995). In study 1, we assessed ambivalence subjectively by asking respon-
dents to rate the extent to which they held mixed feelings about the issue of
women’s rights (1 = very certain of opinion, 3 = somewhat certain, 5 = somewhat
mixed feelings, 7 = very mixed feelings).
Attitude accessibility. Attitude accessibility refers to the ease or quickness
with which an attitude can be retrieved from memory. The time required for
respondents to answer the following attitudinal inquiry was recorded and served

5 Rather than using thermometer ratings, we would have preferred to ask respondents to rate the extent
to which they identified with or cared about each group (see Boninger et al., 1995). However, time
constraints on the survey precluded this possibility.
Involvement and Accessibility 91

Figure 2. Relations among involvement, ambivalence, extremity, and accessibility

(all paths are p < .01).

as a measure of attitude accessibility: “Please tell me whether you think the phrase
‘women’s rights’ represents something ‘good’ or something ‘bad’.” Reaction time
(in hundredths of a second) was computed by taking the elapsed time between the
end of the interviewer’s inquiry and the beginning of the response. Respondents
were asked to respond as quickly as possible without sacrificing accuracy.

Results and Discussion

To evaluate our prediction that attitude extremity and ambivalence mediate

the relationship between involvement and attitude accessibility, we conducted an
observed variable path analysis (Figure 2) using LISREL 8 (Jöreskog & Sörbom,
1993). This model provided a good fit to the data (adjusted goodness-of-fit
index = .92, comparative fit index = .91). Specifically, extremity and ambivalence
were a direct function of involvement, and attitude accessibility was a direct
function of extremity and ambivalence (Figure 2). The indirect paths from involve-
ment through extremity and ambivalence to accessibility were significant (βs = .05,
p < .001). However, above and beyond the indirect (mediated) effects, involvement
continued to have a direct effect on accessibility (45.50% of the relationship
between involvement and accessibility was mediated by extremity and ambiva-
lence). This suggests that although these two structural properties of attitudes
constitute one set of mediators of the involvement-accessibility relationship, the
effects are rather small in magnitude, and that additional factors would appear to
be important in explaining this relationship. One possibility is that highly involving
92 Lavine et al.

attitudes are the targets of frequent activation and behavioral action. The frequency
and recency with which an attitude is activated has been shown to influence the
ease with which the attitude can be retrieved from memory (e.g., Fazio et al., 1982,
experiment 3; see also Higgins & King, 1981; Thomsen et al., 1995; for other
possible mediators of the involvement-accessibility mediators, see Krosnick, 1989;
Lavine et al., 1996).
Although we tested an involvement → accessibility model in study 1, alterna-
tive causal flows are possible. For example, people may infer that attitudes that
come easily to mind are personally important to them (for an experimental
demonstration of the accessibility → involvement causal flow, see Roese & Olson,
1994).

Study 2

One of the important theoretical implications of the involvement-accessibility

relationship is that it can provide one explanation for why involving issue attitudes
typically exert stronger effects on candidate evaluation and voting than do rela-
tively uninvolving attitudes (e.g., Krosnick, 1988a; for a review, see Lavine et al.,
1996). If people’s most involving attitudes are used to inform their candidate
appraisals and political decision-making because such attitudes are the most
chronically accessible, we should find that accessibility is greatest for whatever
issues a person judges as most salient, and that accessibility declines as the issues
become less important to that individual. Therefore, instead of determining whether
attitudes are more accessible for a given issue among those who rate the issue as
important than among those who rate that issue as unimportant (as we did in
study 1), we would need to examine the relationship between attitude involvement
and attitude accessibility across a large number of issues within the cognitive
systems of individuals. A more critical test of the role of attitude structure (i.e.,
extremity and ambivalence) in mediating the relationship between attitude involve-
ment and attitude accessibility thus requires an analytic design that allows for the
examination of the relationship between involvement, extremity, ambivalence, and
accessibility within (rather than between) individuals. In study 2, we used this
within-person, between-issue methodology to examine whether the involvement-
accessibility relationship can be explained, at least partially, by attitude extremity
and attitude ambivalence.6

6 We previously reported the direct within-person relationship between involvement and accessibility
using the data from study 2 (Lavine et al., 1996). However, in that report, we did not examine mediators
of that relationship. The purpose of study 2 is to determine the extent to which the involvement-
accessibility relationship can be accounted for by extremity and ambivalence.
Involvement and Accessibility 93

Method

To evaluate the involvement → extremity/ambivalence → accessibility hy-

pothesis within the cognitive systems of individuals, we assessed the importance,
extremity, ambivalence, and accessibility of seven domestic and seven foreign
policy issues.7 The participants were 84 undergraduates at the University of
Minnesota who participated in the study for extra credit. To assess attitude
involvement, we asked participants to rank-order the 14 issues in terms of their
personal importance (1 = least important issue, 14 = most important issue). We
assessed attitude extremity for each issue by taking the absolute deviation from the
midpoint of a 7-point semantic differential scale (scores thus ranged from 0 to 3).
To assess attitude ambivalence, we used 4-point scales to independently assess the
extent to which respondents held positive and negative evaluations toward each of
the 14 issues. For example, to assess the positive aspects of respondents’ attitudes
toward legalized abortion, we asked respondents to consider only their positive
feelings and beliefs about the policy of legalized abortion and to ignore their
negative feelings and beliefs about the policy. Respondents were then asked
whether they were not at all positive toward the issue (0), slightly positive
(1), moderately positive (2), or extremely positive (3) toward the issue. We then
assessed the degree of negative evaluation in the same manner (for a review of
related measures, see Thompson et al., 1995).
Ambivalence scores for each issue were derived using the Griffin formula (see
Thompson et al., 1995). For each of the 14 issues, this index involves averaging
the positive and negative evaluative components and subtracting the absolute
difference between the positive and negative components. This index operational-
izes two key theoretical properties of attitude ambivalence. First, ambivalence
scores increase as the positive and negative components become increasingly
similar in magnitude. Second, scores increase as the two components become more
intense (see Breckler, 1994; Meffert, Lodge & Guge, in press; Thompson et al.,
1995).
Finally, attitude accessibility was measured by latency of response to a direct
attitudinal inquiry. Respondents were presented with verbal representations (e.g.,
“Legalized Abortion”) of each issue on a computer screen and were instructed to
respond to each issue as quickly and as accurately as possible in terms of whether
it represented something “good” or “bad” by pressing one of two keys on the

7 The issues were all on the public agenda during the period of time in which our data were collected
(winter/spring 1991). The domestic issues consisted of flag burning, legalized abortion, affirmative
action, the women’s rights movement, capital punishment, increased taxes, and unemployment. The
foreign policy issues were increased defense spending, the arms race, the strategic defense initiative,
the Persian Gulf war, Palestinian rights, nuclear war, and the former Soviet Union.
94 Lavine et al.

keyboard. The order of presentation of the stimuli was individually randomized for
each respondent.

Results and Discussion

If attitudes are most accessible, extreme, and univalent (i.e., unambivalent)

toward whatever issues are most involving to a person and least accessible,
extreme, and univalent toward whatever issues are judged as least involving to that
person, there should be a within-person linear relationship between involvement
and each of these three attitude properties. For example, if person A views capital
punishment as the most involving issue and welfare spending as the least involving
issue, attitudes toward capital punishment should be more accessible, extreme, and
univalent than attitudes toward welfare spending for person A. If, however, person
B regards welfare spending as more involving than capital punishment, attitudes
toward welfare spending should be more accessible, extreme, and univalent than
attitudes toward capital punishment for person B. Hence, we performed three
14-level repeated-measures trend analyses with the 14 levels of involvement as the
repeated independent variable—one on respondents’ extremity scores, one on their
ambivalence scores, and a third on their accessibility (i.e., response latency) scores.
Thus, extremity, ambivalence, and accessibility scores served as the dependent
variable in each analysis, respectively. Each of these linear trends was significant
and accounted for a sizable portion of the variance (η2 = .37 for extremity, .25 for
ambivalence, and .35 for accessibility; ps < .001). That is, as involvement in-
creased, attitudes became more extreme, univalent, and accessible.
The primary question we wish to answer in study 2 is whether attitude
extremity and ambivalence mediate the relationship between attitude involvement
and attitude accessibility. If so, then the involvement-accessibility relationship
should be substantially attenuated when it is adjusted for extremity and ambiva-
lence. To examine these mediational hypotheses, we recomputed the involvement-
accessibility analysis of variance three times, once controlling for attitude
extremity, once controlling for attitude ambivalence, and a third time controlling
for both of these properties simultaneously.

Table I. Strength of Within-Person Involvement-Accessibility Relationship With Extremity and

Ambivalence as Covariates

Proportion of covariance
Involvement-accessibility relationship F η2 accounted for by covariate

No covariate 41.73 .3485 —

Extremity covariate 16.01 .1971 .4327
Ambivalence covariate 27.02 .3035 .1291
Extremity/ambivalence covariate 12.99 .2095 .3989

Note. All ps < .001.

Involvement and Accessibility 95

The linear effect of extremity accounted for a substantial portion of the

involvement-accessibility relationship (Table I). Specifically, when scores were
controlled for extremity, the (η2 for involvement on accessibility dropped from
.3485 to .1971, or 43.49%). Ambivalence, however, accounted for much less of the
involvement-accessibility relationship. The involvement-accessibility relationship
was attenuated only 12.79% when ambivalence scores were held constant (Table I).
These results—based on a within-persons analysis of attitude structure—are only
partially consistent with those of study 1. As in study 1, the results of study 2 sug-
gest that when people become involved in a political issue, their attitudes become
more accessible because involvement leads people to form more evaluatively
extreme and univalent attitudes. That is, involvement leads to the formation of
attitudes in which the underlying affective and cognitive constituents are predomi-
nantly on one side of the issue. This relative absence of underlying evaluative
conflict produces attitudes that are relatively easy to retrieve from memory.
Although attitude extremity accounted for a substantial portion of the involve-
ment-accessibility covariance, a substantial portion of the covariance (56.51%) was
left unexplained. Again, this suggests that additional factors mediate the involve-
ment-accessibility relationship. One possibility is that other aspects of attitude
structure that are not isomorphic with extremity or ambivalence explain additional
covariance (e.g., interattitudinal consistency, integrative complexity).
One question that studies 1 and 2 cannot address is why involving attitudes
tend to be extreme and univalent. In study 3, we directly addressed the role of
selective information processing in mediating the involvement-extremity/ambiva-
lence relationship.

Study 3

According to social and political psychological models of reasoning, people

are often motivated by the desire to maintain their current perceptions, beliefs, and
attitudes (e.g., Giner-Sorolla & Chaiken, 1997; Kunda, 1990; Liberman & Chaiken,
1992; Lodge & Taber, in press; McGraw et al., 1996; Schatz, Staub, & Lavine,
1999). This is especially likely to be the case when an issue strongly impinges on
one’s tangible goals, values, and social identifications—that is, when an issue
provokes high levels of involvement (see Lodge & Taber, in press). In study 3, we
examined one mechanism through which attitudinal perseverance might occur, and
two structural consequences of that perseverance. Specifically, our intent was to
evaluate whether high levels of involvement in an issue are associated with a
preference for exposure to attitude-congruent information. If this is indeed the case,
then involvement-based differences in selective exposure would provide one
psychological explanation for why involving attitudes are associated with extreme
and univalent attitude structures. Specifically, as a result of disproportionate
exposure to only one set of political arguments, the affective and cognitive
constituents on which the attitude is based (i.e., the considerations; see Zaller, 1992)
96 Lavine et al.

are likely to be predominantly on one side of the issue (either positive or negative).
Attitudes subject to these processes should thus be associated with low levels of
ambivalence and high levels of extremity.
In study 3, we directly tested this idea by assessing respondents’ levels of
involvement in the issue of affirmative action (an issue that has for some time been
the subject of heated debate on college campuses and in society at large). We also
asked participants to rate their interest in reading various articles that endorsed
either a pro– or an anti–affirmative action position. Finally, we assessed the
extremity and ambivalence of respondents’ attitudes toward the issue. We hypothe-
sized that involvement would be associated with both extreme and univalent
attitudes and with a selective exposure effect (favoring congenial information);
moreover, we explored whether the involvement-extremity/ambivalence relation-
ship would be mediated by involvement-based differences in the tendency for a
preference for attitude-congruent information.

Method

One hundred seventy-nine undergraduates at Northern Illinois University

participated in the study for extra credit. Attitudinal involvement was assessed
using six 9-point items. Participants were asked how much they cared about the
issue, how much time they spent thinking about the issue, and how much they were
interested in information about the issue (α = .87). Attitude extremity was as-
sessed by averaging the absolute difference from the midpoint of four 9-point
semantic differential items (bad–good, harmful–beneficial, foolish–wise, and
unnecessary–necessary) (α = .86). Attitude ambivalence was assessed using both
subjective and objective measures (see Lavine, Huff, et al., 1998; Priester & Petty,
1996; Thompson et al., 1995). The subjective items included asking participants
to rate whether their beliefs and feelings about affirmative action were “mixed” or
“all on one side of the issue” and to rate the degree of conflict they felt about the
issue (see Tourangeau, Rasinski, D’Andrade, & Bradburn, 1989). Three sets of
objective unipolar items were used to separately assess the positive and negative
components of participants’ attitudes. One set of items asked participants to
separately rate the degree to which their feelings about affirmative action were
positive (0 = no positive emotions, 3 = extremely positive emotions) and negative
(0 = no negative emotions, 3 = extremely negative emotions) (affective ambiva-
lence). A second set of items asked participants to separately rate the degree to
which their beliefs about affirmative action were positive (0 = not at all beneficial,
3 = extremely beneficial) and negative (0 = not at all harmful, 3 = extremely
harmful) (cognitive ambivalence). The third set of items asked participants to
separately rate their positive (0 = not at all good, 3 = extremely good) and negative
(0 = not at all bad, 3 = extremely bad) evaluations of affirmative action (general
evaluative ambivalence). Ambivalence scores were constructed from each set of
Involvement and Accessibility 97

items using the Griffin formula (the average of the positive and negative compo-
nents minus the absolute value of the difference; for a review of numerical indices
of ambivalence, see Breckler, 1994; Priester & Petty, 1996; Thompson et al., 1995).
Final ambivalence scores were computed by standardizing the three subjective and
four objective measures and averaging them (α = .79).
Selective exposure was assessed by asking participants to rate their interest in
reading two op-ed type articles, one pro– and one anti–affirmative action, where
1 = “I definitely would not like to read this article” and 7 = “I definitely would like
to read this article” (see Eagly & Chaiken, 1993; Frey, 1986; Pomerantz et al., 1995;
Schatz et al., 1999). These two target items were embedded within a larger group
of eight article titles related to other political issues (e.g., welfare, balancing the
budget). To create an index of the extent to which participants favored attitude-
congruent over incongruent articles (i.e., selective exposure to attitudinally con-
genial information), we subtracted the interest rating for the incongruent article
from the interest rating for the congruent article.

Results and Discussion

To assess whether selective exposure mediates the relationship between in-

volvement and attitude structure (i.e., attitude extremity and ambivalence), we
conducted a path analysis using LISREL 8, where involvement was specified to
have both direct (unmediated) and indirect (mediated) effects on attitude extremity
and ambivalence. Higher levels of involvement were significantly associated with

Figure 3. Relations among involvement, selective exposure, ambivalence, and extremity

(†p < .10, *p < .05, **p < .01).
98 Lavine et al.

a preference for exposure to attitude-congruent information over exposure to

incongruent information (Figure 3). Moreover, this selective exposure tendency
was marginally significantly associated with holding more extreme and more
ambivalent attitudes (ps < .09, one-tailed). Finally, the analysis revealed that
involvement has both significant direct effects (ps < .01) and marginally significant
indirect effects (βs = .02, ps < .10, one-tailed). As these results suggest, other
factors beyond selective exposure would appear to be required to fully explain the
involvement-extremity/ambivalence relationship. One factor may be selective
judgment; that is, involvement may lead people to be more critical and rejecting of
attitude-incongruent than congruent information (e.g., Houston & Fazio, 1989;
Lord et al., 1979). Involving attitudes may also become associated with relatively
univalent attitude structures through group polarization and other interpersonal
mechanisms (e.g., Abelson, 1995).

Conclusions

By incorporating the concept of accessibility into research on political behav-

ior, political psychologists have generated important insights into the nature of
candidate evaluation and issue voting (e.g., Krosnick, 1988a; Lau, 1989), judg-
ments of presidential performance (e.g., Iyengar et al., 1984), the role of party
identification (Bassili, 1995), and the nature and structure of political attitudes (e.g.,
Judd et al., 1991; Thomsen et al., 1996; Zaller, 1992). Collectively, this work
amplifies theory and research in social cognition on the role of construct accessi-
bility in moderating social judgment and decision-making (e.g., Higgins, Rholes,
& Jones, 1977; Wyer & Srull, 1981).
We recognize that variation in both contextual and chronic attitude accessibil-
ity contributes to our understanding of the nature and influence of public opinion
and political processes. But we also believe that deeper psychological insights
require a conceptual framework for understanding how attitudes become accessi-
ble. In this article, we sought to test a model of the cognitive and motivational
processes that mediate the relationship between attitude involvement and attitude
accessibility. We focused on the involvement-accessibility relationship for two
reasons: First, the accessibility of political issues reliably varies as a function of
involvement (Krosnick, 1989; Lavine et al., 1996). Second, the theoretical antece-
dents of accessibility are the hallmarks of involving attitudes. Relative to uninvolv-
ing attitudes, involving attitudes are more frequently thought about (Thomsen et
al., 1995), more strongly linked to people’s ongoing needs, goals, and values
(Boninger et al., 1995; Thomsen et al., 1995), and more embedded within a larger
system of other attitudes and beliefs (Lavine, 1994).
We suggest that at least three key processes are responsible for connecting
subjectively important with easily retrievable attitudes. First, as attitudes become
increasingly central to the self-concept, defense-related goals (e.g., attitude main-
tenance) should exert a stronger influence on issue-related information processing.
Involvement and Accessibility 99

Second, defense-related goals are enacted through “motivated reasoning” or selec-

tivity of exposure, elaboration, judgment, and recall (Eagly & Chaiken, 1993;
Lodge & Taber, in press; McGraw et al., 1996). For example, we would expect
political messages to be cognitively elaborated to the extent that they support
favorable positions or refute unfavorable ones. Alternatively, information might
be met with greater criticality, skepticism, and refutation when it opposes rather
than supports preferred positions. For example, Borgida and his colleagues (Bor-
gida & Howard-Pitney, 1983, experiment 2; Howard-Pitney et al., 1986) have
reported that highly involving attitudes tend to be associated with biases in the
processing of issue debates. The third key process connecting involvement to
accessibility consists of the influence of selective processing on structural proper-
ties of the political attitudes that are formed (or changed). As noted above, biased
or selective processing should result in the disproportionate acceptance of either
positive or negative feelings and beliefs about an issue, rendering the attitude
extreme and univalent.
We should, however, acknowledge the piecemeal and suggestive (rather than
conclusive) nature of the present studies. We did not directly measure whether
processing goals (accuracy vs. defense) are directly caused by involvement, nor
did we link processing goals with selective information-processing strategies.
Finally, we examined only one type of “biased” processing: selective exposure. As
discussed earlier, attitudinal defense can also be accomplished through selective
judgment and recall and the selective use of heuristic cues (e.g., Giner-Sorolla &
Chaiken, 1997). These selective processing strategies are likely to prevail when
counterattitudinal information is perceived as being weak and refutable (see Frey,
1986). Future research should incorporate a broader range of selectivity processes
and should link involvement, processing goals and strategies, attitude structure
(i.e., extremity and ambivalence), and attitude accessibility in a single study.
As the model specifies, and as the data from studies 1 and 2 confirm, political
attitudes with extreme and univalent structures are easier to retrieve from memory
than are those with moderate and ambivalent structures. Resolving attitudinal
ambivalence is a time- and resource-consuming operation (see Tourangeau &
Rasinski, 1988); conflict thus tends to reduce attitude accessibility. However, the
mere availability of conflicting attitudinal considerations in long-term memory
does not necessitate a slowdown in response time to a direct attitudinal inquiry.
Even when such “potential” conflict exists, contextual information may dispropor-
tionately activate beliefs and feelings on one side of the issue, rendering the attitude
highly accessible. In other words, attitudes with ambivalent underlying structures
may be accessible under certain conditions. Bassili (1998) has identified an
important moderator of the relationship between attitude ambivalence and attitude
accessibility. Specifically, Bassili argued that for conflictual beliefs to attenuate an
attitude’s accessibility, the positive and negative evaluation of the attitude object
must be simultaneously activated. He showed that ambivalence was linked to
decreased accessibility only when both the positive and negative components of
100 Lavine et al.

the attitude were accessible. Bassili’s (1998) findings comport well with the
constructionist notion that a given attitude object is likely to provoke a range of
attitudinal responses, depending on the current activation values of available
feelings and beliefs (see Lavine, Huff, et al., 1998).
Finally, by explicating the cognitive and motivational processes through which
political attitudes become accessible, we stand to gain a more complete under-
standing of key political psychological phenomena such as agenda setting and
priming (Iyengar & Kinder, 1987) and the nature of public opinion and the survey
response (Lavine, Huff, et al., 1998; Zaller & Feldman, 1992).
Media effects and priming. Iyengar and Kinder’s (1987; Iyengar et al., 1984)
work on the agenda-setting and priming effects of television news indicates that
the news media—through stressing some issues while ignoring others—shape
citizens’ judgments of national priorities and influence the bases of electoral
decision-making and presidential performance evaluation. Iyengar and Kinder
explained their priming effects on the basis of an “accessibility heuristic”—the idea
that political judgments and decisions are based on whatever information happens
to come to mind. Their theoretical model thus specifies that political circumstances
determine what information will be most salient to citizens and therefore what
information drives judgments and decisions. Although Iyengar and Kinder’s (1987;
Iyengar et al., 1984) studies provide direct evidence for the priming effect, they do
not empirically address the processes of how exposure to television news facilitates
priming. Our involvement-accessibility model can be useful in understanding this
process.
According to the agenda-setting aspect of Iyengar and Kinder’s model, media
coverage of political issues determines the public’s views of which problems are
the most important to the country. We would argue that media coverage similarly
influences the public’s judgments of which problems are most important to the self.
That is, media coverage of issues may invoke involvement by linking political
issues to citizens’ self-interests, values, and social identifications. According to our
model, once citizens are involved in an issue, their media-based information-
processing strategies should become more selective; thus, they should manifest a
greater interest in attitude-congruent than incongruent news stories and should be
selectively critical of incongruent stories. This should produce a set of underlying
attitude considerations that are highly consistent with each other, resulting in
attitudes that are extreme and univalent, and thus highly accessible. Future research
might extend agenda-setting effects to examine citizens’ cognitive responses to
attitude-congruent and incongruent news stories and link such responses to the
structure and accessibility of resulting attitudes.
The nature of political attitudes and the survey response. Recent studies have
called into question the traditional view that political attitudes are stable evaluative
constructs that are represented in memory in summary (precomputed) form and are
directly retrieved from memory when a survey response is required (e.g., Lavine,
Huff, et al., 1998; Tourangeau & Rasinski, 1988; Wilson & Hodges, 1992; Zaller
Involvement and Accessibility 101

& Feldman, 1992). The new view is that people do not possess any single attitude
toward an issue, but rather “carry around in their heads a mix of only partially
consistent ideas and considerations” (Zaller & Feldman, 1992, p. 579). When a
survey response is requested, respondents are believed to canvass their memories
for relevant beliefs and feelings, and then integrate them on the spot to select a
survey response. In doing so, people have been shown to oversample from memory
information made temporarily accessible by the prior survey context, resulting in
item context effects (e.g., Tourangeau et al., 1989). Moreover, the summary
judgment (i.e., the attitude) derived from this retrieval/integration process is
believed to decay over time. Thus, attitudes are hypothesized to be temporary rather
than stable constructs, and survey responses are hypothesized to require computa-
tion rather than simple retrieval.
Our model suggests two potentially important limiting conditions on the
“temporary constructs” view of political attitudes and the mass survey response.
First, as shown in studies 1 and 2, attitudes that are highly connected with the self
have underlying structures in which the considerations (beliefs and feelings) are
highly consistent. Thus, the computation (retrieval and integration) of such atti-
tudes should result in less within-person attitude variability over time than for
attitudes not marked by involvement. Hence, involvement might produce stability
in attitudes (see Krosnick, 1988b; Schuman & Presser, 1981) because involving
attitudes are associated with highly consistent underlying considerations.
Second, the survey response process itself may differ qualitatively across
levels of attitude involvement, extremity, ambivalence, and accessibility. Specifi-
cally, although most attitudes are likely to conform (at least in part) to the
attitudes-as-temporary-constructions view, attitudes marked by high levels of
involvement, extremity, and univalence—and thus accessibility—are likely to
conform to the direct-retrieval, attitudes-as-stable-constructs view. In particular,
accessibility should be positively related to the probability that the attitude exists
in precomputed form, thus bypassing the belief retrieval and integration stages of
the survey response process. These ideas are consistent with Lavine, Huff, et al.’s
(1998) recent finding that “strong” (e.g., involving) attitudes are less susceptible
than “weak” attitudes to item context effects in surveys.
An influential political commentator once declared that any attempt to char-
acterize the nature of public opinion is a task not unlike coming to grips with the
Holy Ghost (Key, 1961, p. 8). However, if political psychologists are to provide a
more complete understanding of the formation, structure, and dynamic operation
of public opinion and electoral behavior, we must direct our attention to the
cognitive and motivational processes that make attitudes accessible in memory,
and thus politically consequential.
102 Lavine et al.

ACKNOWLEDGMENTS

We thank Paul Sniderman, Henry Brady, and Phil Tetlock for organizing the
collection of the data from study 1. Portions of this article were presented at the
symposium on Political Psychology at the October 1996 annual meeting of the
Society of Experimental Social Psychology, Amherst, MA, and at the symposium
on Response Latency Measurement in Survey Research at the July 1998 annual
meeting of the International Society of Political Psychology, Montreal. Correspon-
dence concerning this article should be sent to Howard Lavine, Department of
Political Science, State University of New York at Stony Brook, Stony Brook, NY
11794-4392. E-mail: [email protected]

REFERENCES

Abelson, R. P. (1995). Attitude extremity. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength:
Antecedents and consequences (pp. 25–42). Mahwah, NJ: Erlbaum.
Aldrich, J. H., Sullivan, J. L., & Borgida, E. (1989). Foreign affairs and issue voting: Do presidential
candidates waltz before a blind audience? American Political Science Review, 71, 111–130.
Alvarez, R. M., & Brehm, J. (1995). American ambivalence toward abortion policy: A heteroskedastic
probit method for assessing conflicting values. American Journal of Political Science, 39,
1055–1082.
Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78, 171–206.
Bassili, J. N. (1995). On the psychological reality of party identification: Evidence from the accessibility
of voting intentions and of partisan feelings. Political Behavior, 17, 339–358.
Bassili, J. N. (1996). The “how” and “why” of response latency measurement in survey research. In
N. Schwarz & S. Sudman (Eds.), Answering questions: Methodology for determining cognitive
and communicative processes in survey research (pp. 319–346). San Francisco: Jossey-Bass.
Bassili, J. N. (1998, July). Simultaneous accessibility: A prerequisite to heated intrapsychic conflict.
Paper presented at the annual meeting of the International Society of Political Psychology,
Montreal.
Boninger, D., Krosnick, J. A., & Berent, M. K. (1995). Origins of attitude importance: Self-interest,
social identification, and value relevance. Journal of Personality and Social Psychology, 68,
61–80.
Borgida, E., & Howard-Pitney, B. (1983). Personal involvement and the robustness of perceptual
salience effects. Journal of Personality and Social Psychology, 45, 560–570.
Breckler, S. J. (1994). A comparison of numerical indexes for measuring attitude ambivalence.
Educational and Psychological Measurement, 54, 350–365.
Brehm, J. W., & Cohen, A. R. (1962). Explorations in cognitive dissonance. New York: Wiley.
Brock, T. C., & Balloun, J. L. (1967). Behavioral receptivity to dissonant information. Journal of
Personality and Social Psychology, 6, 413–428.
Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space: A
critical review, with emphasis on the separability of positive and negative substrates.
Psychological Bulletin, 115, 401–423.
Cacioppo, J. T., Gardner, W. L., & Berntson, G. G. (1997). Beyond bipolar conceptualizations and
measures: The case of attitudes and evaluative space. Personality and Social Psychology Review,
1, 3–25.
Involvement and Accessibility 103

Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic processing within and
beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought
(pp. 212–252). New York: Guilford.
Crano, W. D. (1995). Attitude strength and vested interest. In R. E. Petty & J. A. Krosnick (Eds.), Attitude
strength: Antecedents and consequences (pp. 131–158). Mahwah, NJ: Erlbaum.
Dovidio, J. F., Evans, N., & Tyler, R. B. (1986). Racial stereotypes: The contents of their cognitive
representations. Journal of Experimental Social Psychology, 22, 22–37.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace
Jovanovich.
Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T. Fiske, &
G. Lindzey (Eds.), The handbook of social psychology (4th ed., vol. 1, pp. 269–322). New York:
Harcourt Brace Jovanovich.
Eagly, A. H., Chen, S., Chaiken, S., & Shaw-Barnes, K. (1999). The impact of attitudes on memory:
An affair to remember. Psychological Bulletin, 125, 64–89.
Fazio, R. H. (1986). How do attitudes guide behavior? In R. M. Sorrentino & E. T. Higgins (Eds.),
Handbook of motivation and cognition: Foundations of social behavior (pp. 204–243). New York:
Guilford.
Fazio, R. H., Chen, J. M., McDonel, E. C., & Sherman, S. J. (1982). Attitude accessibility,
attitude-behavior consistency, and the strength of the object-evaluation association. Journal of
Experimental Social Psychology, 18, 339–357.
Fazio, R. H., & Williams, C. J. (1986). Attitude accessibility as a moderator of the attitude-perception
and attitude-behavior relations: An investigation of the 1984 presidential election. Journal of
Personality and Social Psychology, 51, 505–514.
Feldman, S., & Zaller, J. (1992). Political culture of ambivalence: Ideological responses to the welfare
state. American Journal of Political Science, 36, 268–307.
Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson.
Festinger, L. (1964). Conflict, decision, and dissonance. Stanford, CA: Stanford University Press.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory
and research. Reading, MA: Addison-Wesley.
Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from category-based to
individuating processes: Influences of information and motivation on attention and interpretation.
Advances in Experimental Social Psychology, 23, 1–74.
Freedman, J. L., & Sears, D. O. (1965). Selective exposure. Advances in Experimental Social
Psychology, 2, 57–97.
Frey, D. (1986). Recent research on selective exposure to information. Advances in Experimental Social
Psychology, 19, 41–80.
Giner-Sorolla, R., & Chaiken, S. (1997). Selective use of heuristic and systematic processing under
defense motivation. Personality and Social Psychology Bulletin, 23, 84–97.
Gorn, G. J. (1975). The effects of personal involvement, communication discrepancy, and source
prestige on reactions to communications on separatism. Canadian Journal of Behavioral Science,
7, 369–386.
Greenwald, A. G., & Breckler, S. J. (1985). To whom is the self presented? In B. R. Schlenker (Ed.),
The self and social life (pp. 126–145). New York: McGraw-Hill.
Higgins, E. T., & King, G. (1981). Accessibility of social constructs: Information-processing
consequences of individual and contextual variability. In N. Cantor & J. Kihlstrom (Eds.),
Personality, cognition, and social interaction (pp. 69–122). Hillsdale, NJ: Erlbaum.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility and impression formation.
Journal of Experimental Social Psychology, 13, 141–154.
Hochschild, J. L. (1981). What’s fair? American beliefs about distributive justice. Cambridge, MA:
Harvard University Press.
104 Lavine et al.

Houston, D. A., & Fazio, R. H. (1989). Biased processing as a function of attitude accessibility: Making
objective judgments subjectively. Social Cognition, 7, 51–66.
Howard-Pitney, B., Borgida, E., & Omoto, A. M. (1986). Personal involvement: An examination of
processing differences. Social Cognition, 4, 39–57.
Iyengar, S., & Kinder, D. R. (1987). News that matters: Television and American opinion. Chicago:
University of Chicago Press.
Iyengar, S., Kinder, D. R., Peters, M. D., & Krosnick, J. A. (1984). The evening news and presidential
evaluations. Journal of Personality and Social Psychology, 46, 778–787.
James, W. (1890). The principles of psychology. New York: Holt, Rinehart and Winston.
Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-analysis.
Psychological Bulletin, 106, 290–314.
Jöreskog, K., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS
command language. Hillsdale, NJ: Erlbaum.
Judd, C. M., Drake, R. A., Downing, J. W., & Krosnick, J. A. (1991). Some dynamic properties of
attitude structures: Context-induced response facilitation and polarization. Journal of Personality
and Social Psychology, 60, 193–202.
Judd, C. M., & Krosnick, J. A. (1989). The structural bases of consistency among political attitudes:
Effects of political expertise and attitude importance. In A. R. Pratkanis, S. J. Breckler, &
A. G. Greenwald (Eds.), Attitude structure and function (pp. 99–128). Hillsdale, NJ: Erlbaum.
Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24,
163–204.
Key, V. O., Jr. (1961). Public opinion and American democracy. New York: Knopf.
Krosnick, J. A. (1988a). The role of attitude importance in social evaluation: A study of policy
preferences, presidential candidate evaluations, and voting behavior. Journal of Personality and
Social Psychology, 55, 196–210.
Krosnick, J. A. (1988b). Attitude importance and attitude change. Journal of Experimental Social
Psychology, 24, 240–255.
Krosnick, J. A. (1989). Attitude importance and attitude accessibility. Personality and Social
Psychology Bulletin, 15, 297–308.
Krosnick, J. A., Boninger, D. S., Chuang, Y. C., Berent, M. K., & Carnot, C. G. (1993). Attitude strength:
One construct or many related constructs? Journal of Personality and Social Psychology, 65,
1132–1151.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480–498.
Lau, R. (1989). Construct accessibility and electoral choice. Political Behavior, 11, 5–32.
Lavine, H. (1994). The influence of attitudinal involvement on inter-attitudinal consistency and attitude
change (doctoral dissertation, University of Minnesota, 1994). Dissertation Abstracts
International, 56, 566.
Lavine, H. (1997). Reaction time: A methodological tool for political psychology? Political
Psychologist, 2, 3–6.
Lavine, H., Huff, J. W., Wagner, S. H., & Sweeney, D. (1998). The moderating influence of attitude
strength on the susceptibility to context effects in attitude surveys. Journal of Personality and
Social Psychology, 75, 359–373.
Lavine, H., Sullivan, J. L., Borgida, E., & Thomsen, C. J. (1996). The relationship of national and
personal issue salience to attitude accessibility on foreign and domestic policy issues. Political
Psychology, 17, 293–316.
Lavine, H., Thomsen, C. J., & Gonzales, M. H. (1997). The development of interattitudinal consistency:
The shared-consequences model. Journal of Personality and Social Psychology, 72, 735–749.
Lavine, H., Thomsen, C. J., Zanna, M. P., & Borgida, E. (1998). On the primacy of affect in the
determination of attitudes and behavior: The moderating influence of affective-cognitive
ambivalence. Journal of Experimental Social Psychology, 34, 398–421.
Involvement and Accessibility 105

Liberman, A., & Chaiken, S. (1992). Defensive processing of personally relevant health messages.
Personality and Social Psychology Bulletin, 18, 669–679.
Lodge, M., McGraw, K., & Stroh, P. (1989). An impression-driven model of candidate evaluation.
American Political Science Review, 83, 399–419.
Lodge, M., & Taber, C. (in press). Three steps toward a theory of motivated political reasoning. In
A. Lupia, M. McCubbin, & S. Popkin (Eds.), Elements of political reason: Understanding and
expanding the limits of rationality. New York: Cambridge University Press.
Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimilation and attitude polarization: The effects
of prior theories on subsequently considered evidence. Journal of Personality and Social
Psychology, 37, 2098–2109.
McGraw, K. M., Fischle, M., Stenner, K., & Lodge, M. (1996). What’s in a word: Bias in trait
descriptions of political leaders. Political Behavior, 18, 263–287.
Meyer, D. E., & Schvaneveldt, R. W. (1971). Facilitation in recognizing pairs of words: Evidence of a
dependence between retrieval operations. Journal of Experimental Psychology, 90, 227–234.
Neisser, U. (1976). Cognition and reality: Principles and implications of cognitive psychology. San
Francisco: Freeman.
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. Advances in
Experimental Social Psychology, 19, 123–205.
Pomerantz, E. M., Chaiken, S., & Tordesillas, R. (1995). Attitude strength and resistance processes.
Journal of Personality and Social Psychology, 69, 408–419.
Priester, J. R., & Petty, R. E. (1996). The gradual threshold model of ambivalence: Relating the positive
and negative bases of attitudes to subjective ambivalence. Journal of Personality and Social
Psychology, 71, 431–449.
Rahn, W. M., Aldrich, J. H., Borgida, E., & Sullivan, J. L. (1990). A social-cognitive model of candidate
appraisal. In J. A. Ferejohn & J. H. Kuklinski (Eds.), Information and democratic process
(pp. 136–159). Urbana, IL: University of Illinois Press.
Rhine, R. J., & Severance, L. J. (1970). Ego-involvement, discrepancy, source credibility, and attitude
change. Journal of Personality and Social Psychology, 16, 175–190.
Roese, N. J., & Olson, J. M. (1994). Attitude importance as a function of repeated attitude expression.
Journal of Experimental Social Psychology, 30, 39–51.
Schatz, R. T., Staub, E., & Lavine, H. (1999). On the varieties of national attachment: Blind versus
constructive patriotism. Political Psychology, 20, 151–174.
Schuman, H., & Presser, S. (1981). Questions and answers in attitude surveys: Experiments on question
form, wording, and context. San Diego, CA: Academic Press.
Schwarz, N., Frey, D., & Kumpf, M. (1980). Interactive effects of writing and reading a persuasive
essay on attitude change and selective exposure. Journal of Experimental Social Psychology, 16,
1–17.
Sherif, M., & Hovland, C. I. (1961). Social Judgment: Assimilation and contrast effects in
communication and attitude change. New Haven, CT: Yale University Press.
Smith, E. R., & Miller, F. D. (1983). Mediation among attributional inferences and comprehension
processes: Initial findings and a general method. Journal of Personality and Social Psychology,
44, 492–505.
Smith, M. B., Bruner, J. S., & White, R. W. (1956). Opinions and personality. New York: Wiley.
Tetlock, P. E. (1986). A value pluralism model of ideological reasoning. Journal of Personality and
Social Psychology, 50, 819–827.
Thompson, M. M., Zanna, M. P., & Griffin, D. W. (1995). Let’s not be indifferent about (attitudinal)
ambivalence. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and
consequences (pp. 361–386). Mahwah, NJ: Erlbaum.
106 Lavine et al.

Thomsen, C. J., Borgida, E., & Lavine, H. (1995). The causes and consequences of personal
involvement. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and
consequences (pp. 191–214). Mahwah, NJ: Erlbaum.
Thomsen, C. J., Lavine, H., & Kounios, J. (1996). Social value and attitude concepts in semantic
memory: Relational structure, concept strength, and the fan effect. Social Cognition, 14, 191–225.
Tourangeau, R., & Rasinski, K. A. (1988). Cognitive processes under context effects in attitude
measurement. Psychological Bulletin, 103, 299–314.
Tourangeau, R., Rasinski, K. A., D’Andrade, R., & Bradburn, N. (1989). Belief accessibility and context
effects in attitude measurement. Journal of Experimental Social Psychology, 25, 401–421.
Wyer, R. S., Jr., & Srull, T. K. (1981). Category accessibility: Some theoretical and empirical issues
concerning the processing of social stimulus information. In E. T. Higgins, C. P. Herman, &
M. P. Zanna (Eds.), Social cognition: The Ontario symposium (vol. 1, pp. 161–198). Hillsdale,
NJ: Erlbaum.
Zaller, J. R. (1992). The nature and origins of mass opinion. New York: Cambridge University Press.
Zaller, J. R., & Feldman, S. (1992). A simple theory of the survey response: Answering questions versus
revealing preferences. American Journal of Political Science, 36, 579–616.
Zuwerink, J. R., & Devine, P. G. (1996). Attitude importance and resistance to persuasion: It’s not just
the thought that counts. Journal of Personality and Social Psychology, 70, 931–944.